►
From YouTube: Ethereum Core Devs Meeting #71 [2019-09-20]
Description
A
All
right
good
morning,
everybody
and
welcome
to
edition
71
of
the
core
developer
meetings,
we'll
jump
right
in
to
issue
or
agenda
item
number
one.
The
Istanbul
related
client
updates
I'm
opening
up
Martin's
tracker
here
and
it
looks
like
we
have
some
new
stuff.
We
can
just
go
through
the
clients
and
see
what's
new
on
each
one,
we'll
start
with
death
and
Martin.
Oh.
B
C
So
not
tough
money
changes
since
the
last
time,
I
think
his
merits.
We
updated
the
block
numbers.
We've
been
running
the
deaths
that
we've
been
running
so
far.
All
green
and
I
think
we'll
be
testing
the
remaining
batches
or
some
releases
that
are
already
marked
as
Istanbul.
But
after
we
run
all
the
tests,
we'll
probably
do
some
announcements
for
people,
not
the
final
duration
for
Istanbul.
D
A
E
F
F
This
will
be
a
one
point.
Two
point:
four
release
would
be
the
first
one
under
the
face.
You
name
one.
Two
three
has
all
the
code
in
that's
just
not
plugged
into
the
to
the
built
in
adjacent
files.
We
have
to
the
Genesis,
so
you
can
run
either
in
a
test.
So
we'll
have
a
formal
release
that
just
management,
that's
a
bit
on
Monday,
Australian
time,
I.
Think.
F
A
A
A
We
also
had
a
call
with
Gorley
and
theirs
was
that
was
decided
on
that
call
to
be
one
five,
six
one,
six,
five
one
for
October
30th.
If
you
go
to
the
agenda,
it's
the
first
comment
offer
he
left
a
list.
Rinkeby
block
is
gonna,
be
five.
Four
three:
four:
five,
three
four
five
and
november
that
was
a
mouthful
and
Kovan
is
in
december
with
one
four
one,
one
one
four
one.
A
So
we
successfully
got
all
palindromes.
We
really
know
what's
important
in
this
space,
so
the
Blake
to
be
EIP,
update,
I've
been
paying
some
people
about
that
cuz
we
were
wanting
to
merge
it
and
I
know.
Accent
was
talking
to
some
people
I
think,
because
there
was
some
confusion
about
what
the
latest
stuff
was
and
I
think.
The
latest
thing
was
that
just
that
Peter
had
but
accent
can
you
provide
any
updates
on
that?
A
B
A
A
H
H
B
H
B
H
B
I
think,
as
far
as
I
know,
yes,
the
consensus,
the
official
consensus
tests,
which
there
aren't
that
many
they
are
I,
think
Dmitry
is
using
death
to
generate
them.
Apart
from
that,
I
have
generated
so
I
released
I,
don't
continuous,
passing
gas
purchase
30
with
quite
a
lot
of
invocations
of
a
slogan.
A
stories
from
various
places
called
calls.
:
calls
and
calls
recursive
stuff
with
gas
as
the
base
and
and
get
is
highly
consistent
with
parroting
over
7,000,000
test
cases.
B
Yes
and
I
think
some
comments
yesterday
that
it
was
no
business
trip,
but
with
address
address,
the
concerns
I
haven't
actually
looked
if
he
has
time
to
do
that
did
ya
if,
if
anyone
is
listening
with
client
developer,
if
you
know
that
yeah
I
mean
speak
up,
if
you
think
it
sounds
odd,
but
I'm,
assuming
that
all
of
you
already
put
that
condition.
As
the
first
thing
you
do.
C
H
So
the
other
set
of
questions
on
the
PR
forests
were
mostly
about
the
mostly
about
two
things:
one
I
think
a
while
ago
it
was
agreed
that
2,200
should
be
self-contained.
It
should
have
all
the
details
needed
it
shouldn't
just
say
that
you
know
refer
to
1283
4x
and
refer
to
1706
4y,
rather
than
that
it
should
be
self-contained,
but
it
still
has
committees,
alike,
just
minor
I,
guess
e
IP
details
that
is
still
refers
to
them,
as
required
ones
and
I.
H
Think
some
sections
may
not
be
self-contained
I'm,
not
sure
if
I
still
won
debt
or
not,
but
I
just
wanted
to
raise
it
and
the
second
question
people
had
there
is
compared
to
1283.
The
biggest
change
was
to
also
reflect
the
s
load,
gas
change,
but
that
gas
change
isn't
reflected
in
the
in
the
appendix
and
some
other
parts
and
some
measurements
or
tables
in
the
EIP.
It
still
uses
the
the
old
s
load,
cost
I
think
these
were
the
two
major
things.
People
were
confused
about.
A
A
G
A
H
Think
the
the
Dec
pre-compile
AIP
just
received
a
pull
request
to
to
move
it
to
final,
and
maybe
this
would
be
a
good
time
to
briefly
cover
what
would
be
the
process
for
being
final
and
I.
Think
the
short
answer
to
that
is
AIP
one
defines
it,
but
I
don't
think
it
defines
the
process
in
a
really
good
way
for
Corey
ip's,
but
it
says
that
they
should
move
through
this
last
call
process
first
and
and
then
they
can
be
adopted.
H
Maybe
we
don't
really
for
all
these
different
processes
to
take
place
now.
So
maybe
we
could
just
agree
on
this
call
for
some
of
them
to
be
moved
to
final
and
I
would
propose
that
ICI
pre-compiled
and
the
coal
data.
Every
pricing
should
definitely
be
moved
to
final,
because
they
they
seem
unlikely
to
be
able
any
changes
or
I
never
seen
anybody
ask
questions.
We
complain
about
them
and
Martin
also
proposed
to
move
1884
to
final.
So
maybe
we
could
just
take
a
vote
on
this
call.
A
Okay,
does
anyone
have
I
mean
I'm,
not
anticipating
in
the
opposition,
to
making
those
final
but
and
also
I?
Think
Matt
asked
that
1108
get
moved
to
final
as
well
and
that
one
would
be
set
the
alt,
that's
the
alt,
be
in
128
reduction.
I,
don't
think
there
was
anything
left
on
that
I
think
it
can
go
to
final.
I
D
H
D
A
A
A
A
The
fact
that
it
was
updated
or
that
there
was
a
suggestion
that
was
removed
during
the
process
of
coming
up
with
what
EIP
1344
was
so
I,
don't
fully
understand
it,
but
the
I
think
parity
and
Trinity
should
just
watch
out
for
this
update.
It's
kind
of
what
I'm,
reading
and
Brian
couldn't
be
on.
This
call,
but
yeah
any
other
clients.
Just
look
over
that
and
make
sure
you're
following
the
EIP
correctly
yeah.
B
B
A
J
I
mean
I
just
wanted
a
quick
summary.
Basically
I.
We
looked
at
actually
taking
everything
out
of
DRC
and
embedding
it
first
in
the
IP
and
the
issue
that
I
had
is
that
I?
Actually
over
specs
VIP,
a
few
of
us
are
using
implementations
that
do
not
follow
the
same,
for
example
Sigma
or
V,
but
but
the
do
match
the
tests
provided
in
the
EAP,
so
that
was
my
main
I
mean
pushback
when
we
actually
went
to
update
the
IP.
J
The
other
thing
we
did
is
I
just
did
my
best
to
make
sure
that
was
clear.
That
was
Blake,
which
included
Blake
to
be
rather
than
rather
than
just
plate
to
be,
and
also
that
it's
blinked
to
64-bit
variants
in
the
IP,
so
I
know
that
it's
probably
not
perfect,
but
I
hope
that
addresses
the
nominator
concerns.
A
I'm
trying
to
see,
if
there's
anything
else
on
here,
we'll
do
testing
updates
in
a
little
bit
I'll
make
sure
we
get
those
in
well.
Actually
it's
not
on
the
agenda
at
all.
Do
we
want
to
just
go
over
testing
updates?
Now,
that's
something!
That's
gonna
be
interesting.
I
know
there
was
some
fuzzing
stuff
and
hive
updates.
I
B
B
K
B
There
is
also
updates
happening
regarding
the
Lib
processor.
If
you
recall,
we
have
one
set
of
post
testing
that
does
not
use
introspection
and
doesn't
actually
see
how
much
it
covers
and
that's
the
JSON
based
and
the
other
one
is
a
little
faster
which
actually
does
coverage
measurements
during
executions
and
then
modifies
the
input
state
and
the
data
to
get
more
coverage.
That's
the
cursor
and
it's
running
gasps
versus
and
we're
hoping
to
have
it
up
yeah
during
the
weekend
or
early
next
week,
yeah
we
have
a
new
server
for
hive.
B
B
A
K
Sure
I'll
just
keep
it
short,
they're
still
working
out
most
of
the
details,
but
I
dropped
a
link
to
what
we
did
with
the
if
you're
in
classic
test
net
it's.
It
was
just
a
temporary,
publicly
joinable
test
net
that
we
post
and
basically
proving
that
the
hard
for
clients
are
gonna
work
as
planned.
It's
it's
well.
What
we
imagine
is
like
a
step
before
doing
a
public
test
net
like
switching
over
Robson
or
Gurley.
We
can
do
something
here,
prove
our
assumptions
or
test
our
assumptions
and
well.
K
What
we
can
also
do
with
the
white
plug
Genesis
platform
is
what
we're
doing
with
each
to
doing.
Multi
client
hosted
test
nets,
and
then
we
can
also
run
specific
contracts
and
make
sure
they're
gonna
work
specifically
in
the
case
of
like
1884.
We
can
test
when
they're
gonna
break
how
they're
gonna
break
and
and
just
once
we
get
a
certain
amount
of
data.
We
can
dump
that
out
and
do
analysis
on
it,
and
then
we
can
iterate
on
that
run.
K
Different
tests,
nuts
iteratively
one
after
the
other-
and
it's
sort
of
meant
to
be
a
quick,
quick
and
dirty
process.
At
least
for
this,
what
we're
planning
we'll
have
more
details
coming
out
soon
about
the
specifics.
How
many
know
is
how
long
it's
gonna
run
when
we'll
be
publicly
available,
but
that's
the
general
idea.
We
think
it'll
be
an
interesting
experiment
to
add
onto
and
hopefully
supplement,
the
other
public
test
nets.
B
K
Interesting
I
think
they're
gonna
be
clean,
but
the
way
that
the
way
that
we
have
done
in
the
past
is
there's
a
genesis
file
that
instructs
how
to
what
sort
of
information
to
supply.
So
that
might
be
something
we
could
do
in
the
future.
Like
kickstart,
it
with
specific
state
and
specific
parameters
like
like
that.
So.
B
K
C
C
K
One
of
the
other
things
that
we've
built
into
the
product
is
that
you
can,
for
example,
preload
a
bunch
of
wallets
or
set
up
accounts
beforehand
in
a
yellow
file,
for
example.
So
what
we're
trying
to
do
is
just
automate
this
process
of
kick-starting
and
and
abstracting
a
little
bit
of
this
away.
Yeah.
K
Now,
it's
just
CLI
so,
but
in
the
future
it
would
be
just
like
a
web,
a
web
UI
and
you
just
click,
hey
I,
want
a
test
net
of
30
nodes
and
I
want
to
run
this
specific
test
case
and
then
we'll
spin
it
up
for
you,
and
then
you
can
introduce
part
of
what
we
also
do
is
introducing
like
adversarial
conditions
like
you
can
pump
up
the
latency
you
see.
If
things
are
still
gonna
come
to
consensus,
you
can
decrease
bandwidth,
have
network
partition
and
see
if
the
network
comes
back
to
consensus.
B
C
K
K
A
A
Let's
see
next
up,
I
just
wanted
to
give
there's
a
link
on
item
number
four
for
the
etherium
roadmap:
2020
a
community
discussion
at
DEFCON
five
day,
one
so
we're
not
this
year
at
Def,
Con,
there's
not
going
to
be
day
zero.
One.
Two
three!
It's
gonna
be
day
one
day
two
day
three
and
day,
four
so
for
day,
one
there's
not
gonna
be
a
main
stage
that
has
speakers
on
it.
It's
gonna
be
divided
by
partitions,
where
there's
going
to
be
different
workshops,
seminars
and
groups
held
this
etherium
roadmap.
A
2020
community
discussion
is
going
to
be
on
the
8th
from
1
to
5
and
the
convention
room
that
has
a
200
person
capacity.
It's
going
to
talk
about
the
technical
roadmaps
for
eighth
1/8
in
the
transition
from
e1
to
e2.
It
has
the
format
on
here
the
schedule
and
it
ends
with
working
groups.
It
starts
with
an
overview
and
then
the
eighth
one
roadmap
I'm
doing
a
talk
on
8th
one
X,
so
I'll
probably
be
the
main
point
on
the
etherium
1
roadmap,
unless
I
get
pulled
into
something
and
I
have
to
find
someone
else.
A
One
more
thing
to
note:
there
is
a
link
to
the
full
schedule
of
community
sessions
for
September,
8th
under
schedule.
It
just
says
see
the
full
schedule.
So
that's
pretty
interesting
to
see
all
the
stuff
The
Magicians
have
put
together
this
year
and
yeah.
Just
let's
see
oh
man,
this
is
editable.
Oof
I
need
to
get
locked.
This
I
need
to
get
someone
to
lock
this
down.
Anyways.
A
Yeah
so
feel
free
to
join,
that
and
it'll
be
oh
yeah.
It
needs
to
be
October,
8th,
good
call,
whoever
just
edited
that
yeah
so
feel
free
to
join
and
I'll.
Probably
if
there's
another
quartet
meeting
before
Def
Con,
which
we'll
discuss
in
a
little
bit
I'll
remind
you
all
of
this
or
throw
it
on
the
Chordettes
channel,
and
if
you
want
to
help
this,
especially
if
you
have
some
knowledge
that
might
be
really
good
for
this.
A
Please
participate
join
in
there's
a
few
people,
especially
if
you
all
know
Jamie
Jamie,
Pitts
he's
one
of
the
organizers.
So
as
a
net
I
can
not
pronounce
her
last
name
recall:
vaga
I'm.
Can
anyone
pronounce
that
last
name
either
way
it's
a
net
and
they
are
part
of
the
people?
Organizing
this
I
think
also
Maria.
Paula
I
know
that
she's
at
least
the
emcee.
A
So
that's
that
for
that
one,
both
Prague
pal
audits
have
been
released.
They've
been
released
for
a
bit
because
it
happened
right
after
the
last
core
dev
meeting,
so
it
feels
like
forever
ago.
But
if
you
all
want
to
look
at
that,
you
can
just
click
the
link
and
there
is
a
blog
post
from
the
cat
herders
that
has
links
to
both
the
least
authority.
A
Software
audit
and
the
Bob
rau,
Hardware
art
audit
and,
let's
see
Oh
Matt,
just
posted
a
cool
comment
about
having
a
bunch
of
kubernetes
infrastructure
for
quick
tests
nets
that
they
might
open
source
and,
if
someone's
interested
in
that
to
reach
out
to
them,
was
there
anything
else
on
that
Matt
or
just
simple.
Oh.
A
A
A
Okay
review
the
previous
decisions
made
and
action
items
for
call
70,
that's
one's
pretty
light.
Leave
Blake
as
like
to
be
keeping
it
fixed
to
12
rounds,
was
decision.
70.1
I
think
that
might
have
been
tweaked
a
little
bit
because
one
of
the
action
items
was
that
Alex
was
gonna.
Talk
to
the
EIP
152
champions
to
discuss
the
concerns,
so
I
think
some
of
that
might
have
changed.
B
A
A
Great
can
someone
who's
writing
notes,
take
note
to
edit
that
decision
within
the
quartet
meeting.
Seventy
summary
notes:
you
can
just
do
a
PR
on
that
any
of
the
cat
herders
in
here
who
want
to
just
change
that
decision.
So
we
have
it
on
record,
but
that's
not
what
happened
thanks
decision?
Seventy
point
to
the
fork
of
Rob's
intestine
is
set
for
the
second
of
October
and
the
block
corresponds
to
that
date
will
be
selected.
That's
happened
and
then
I
already
talked
about
action.
A
70.1
with
Alex
and
AIP
152
champions
action,
70.2,
where
the
cat
herders
creating
a
communication
that
highlights
the
concerns
around
1884
and
provides
it
back
to
the
core
devs
for
a
review.
So
once
we
have
it
at
a
at
a
point
where
we
can
take
comments
on
it,
we're
going
to
put
it
in
the
awkward
dev
Channel
for
a
little
bit
and
make
sure
everything's
good
on
there.
A
L
A
B
M
L
F
And
you
know
the
stuff
that
they're
doing
at
the
beginning
at
the
end
of
the
process,
and
it
doesn't
really
have
any
security
impact
and
that's
they
came
forward
with
them.
That
satisfies
my
concern.
That's
what
I
was
thinking,
because
it's
such
a
deep
in
the
weeds
thing.
It's
not
going
to
subject
an
extension
attack
when
you're
immediately
going
after
another
round,
with
all
the
details
on
it
and
finishing
it
out,
pretty.
A
A
F
N
K
F
Was
the
extent
of
the
interaction
was
the
Twitter
things
they're
kind
of
hard
to
get
you
get
ahold
of
so
I
copied
it
into
the
I
copied
into
the
prog
power
review
room,
but
the
chat
log
was
kept
rolling
and
scrolling
with
all
the
concerns
about
it.
They
were
being
brought
up
again,
so
it
is
in
the
log
if
we
can
find
it.
I
can
post
it
again.
F
A
A
B
I'm
not
sure
I
understood
the
question,
but
I
mean
profile
has
already
been
merged
into
parity
and
it's
activatable
and
I
guess
we
could
do
the
same
put
gas,
but
the
thing
is
we're
not
sure.
If
we're
gonna
go
with
the
same
version,
that
parity
has,
which
is
I,
think
it
is
or
if
we
should
aim
for
Oh
9
3,
it
changes
the
parameter
that
how
often
you
need
to
recompile
the
keep
your
kernel
I'm,
not
sure,
I
understood
what
the
question
was:
I.
K
Mean
it
gets
it's
broader
than
the
technical
implementation
it.
It
starts
to
touch
on,
or
at
least
what
I've
seen
brought
up.
Is
that
it's
a
more
nuanced
conversation
about
like
the
transition
from
eath
1
to
e
2,
and
how
having
frog
power
could
be
sort
of
a
distillation
to
interrupt
the
deposit
contract
or
it's
sort
of
like
keeping
it
in
our
back
pocket
to
be
used
later
on,
if
necessary.
B
B
A
K
A
Yeah
I
think
it's
personally
I
think
it's
too
early
for
the
alternatives.
I
mean
it
feels
like
everything's
like
been
in
forever
ago,
but
really
the
everyone's
screaming
about.
It
was
like
only
after
bubs
our
audit,
which
has
been
released
in
the
last
two
weeks,
so
yeah
or
if
there
was
screaming
about
it
before
it
kind
of
died
down
for
a
while.
So
we
didn't
really
get
any
major
I
guess:
I
didn't
collect
that
much
sentiment
from
the
community
enough
to
bring
it
up
as
an
item
yet
up
until
now
and
now
I'm
collecting
sentiment.
K
B
B
Double
chains
so
that
they
do
some
double
stands
and
something
happens,
and
it
doesn't
matter
really
what
but,
certainly
okay,
we
we
need
to
activate
proper
and
the
problem.
If
that
happens
is
I,
think
that's
which
cannot
be
done
in
a
day
or
an
hour.
It
can
be
done
my
way
over
two
weeks,
and
once
it's
been
decide
to
do
that,
then
all
the
existing
minor
infrastructure
know
that
hey,
there's,
we're
gonna
be
kicked
out
and
they're
going
to
play,
do
whatever
they
want
with
a
chain
because
they're
their
investment
is
going
to
be
worthless.
B
F
Was
just
going
to
point
out,
some
of
the
political
discussion
has
been
attacks
on
small
core
devs
as
well,
so
the
operatives
are
aware
of
some
of
the
politics
going
on
with
it.
It's
not
like
we're
sticking
our
heads
in
the
sand,
we're
just
not
engaging
in
a
political
process
because
we
don't
see
the
constructive
nature
of
what's
been
going
on.
A
M
A
Okay,
thanks
Trent
James
was
there
anything
else,
nope,
cool,
okay,
so
I'll
do
that
sentiment,
collection,
we'll,
keep
our
ears
open
and
do
to
the
new,
mostly
political
arguments,
a
few
technical
arguments
that
kind
of
thing
and
yeah
try
to
be
civil
about
this.
Everyone
listening
in
because
if
you're
loud
and
angry
I'm
gonna
be
less
likely
to
take
you
seriously.
So
that's
me
personally.
K
F
That's
I
think
an
argument
Oh
support
when
we
did
a
minor
pool
with
70%
participation.
There
was
100%
in
support
of
it.
Not
a
single
minor
took
the
time
to
do
a
negative
vote.
If
there
was
a
single
minor
that
took
the
time
for
the
neighborhood
in
I
would
put
more
weight
in
that
argument.
To
be
honest,
just
yeah.
I
That
also
I
checked
for
fun
yesterday
in
a
very
non-scientific
way,
but
there
still
and
he
sipped
amount
of
minors,
signaling
frog,
pal,
yes
and
their
blocks,
so
I'm
not
sure
like
what
percentage
it
is
and
and
whatnot,
but
just
like
randomly
scrolling
through
the
blocks.
It
seems
like
the
minors
who
are
voting
yes,
like
just
did
not
remove
those
votes,
so
it
might
even
still
be
able
to
capture
that
that
sentiment.
L
A
Yeah
that
actually
did
happen
almost
and
that
one
was
actually
like
had
code
behind
it.
So
if
I
see
anything
from
I
will
mention
it
if,
in
my
sentiment,
collection
if
I
see
any
true
attempts
or
plans
or
preparations
for
a
tight
like
a
chain
split
but
I
mean
for
Kings
that
one
of
the
goods
great
things
about
block
chains.
In
my
opinion,
if
you
don't
agree
with
something
you
can
leave,
anybody
else
go
ahead.
A
Yeah
one
final
thing:
we
do
try
to
keep
politics
out
of
the
core.
Devs
calls
it
is
inevitable,
sometimes
and
I.
Think
most
of
us
have
addressed
that
fact
that
it
can
be
inevitable,
since
we're
kind
of
like
Trent
mentioned
somewhat
like
gatekeepers
in
a
way
because
we're
the
ones
who
put
the
code
in
for
everybody
who's
running
the
code,
so
yeah
that's
unfortunate
and
then
Martin.
You
were
mentioning
one
thing
earlier
about
how
gets
on
nine
point
three
right,
I
think
actually
guests.
On
nine
point:
two:
according
to
Andrea
Len
Frenchie.
F
A
Perfect:
okay,
if
there's
nothing
else,
oh,
we
need
to
figure
out
if
we're
meeting
again
like
when
cuz
it's
kind
of
close
to
DEFCON
in
two
weeks,
people
might
be
arriving
in
Japan
at
that
time
and
not
having
very
solid
internet
I've
heard
who's
been
to
Japan.
Is
the
Wi-Fi
really
not
that
prominent
in
Japan
like
at
in
places
or
is
that,
like
a
lie
or
like
something
I'm
reading,
that's
old,
like
news.
A
A
A
Okay,
Tim
will
run
the
meeting
in
two
weeks.
If
there's
really
not
much
on
the
agenda,
we
can
just
cancel
it
change
our
minds
if
people
bring
that
up
on
the
core,
deaf
chat,
all
right
thanks
for
sticking
around
everybody
meeting
in
two
weeks
same
time,
Tim
we'll
be
running
that
and
thank
y'all
for
your
time.
Cheers
thanks.